Volume 37, Issue 2, Pages 249-261 (January 2003) VRILLE Feeds Back to Control Circadian Transcription of Clock in the Drosophila Circadian Oscillator  Nicholas R.J. Glossop, Jerry H. Houl, Hao Zheng, Fanny S. Ng, Scott M. Dudek, Paul E. Hardin  Neuron  Volume 37, Issue 2, Pages 249-261 (January 2003) DOI: 10.1016/S0896-6273(03)00002-3

Figure 1 VRI Protein Cycles in Antiphase with Clk mRNA (A) Western blot analysis of untreated, phosphatase-treated, or phosphatase plus phosphatase inhibitor-treated total proteins extracted from the heads of wild-type and cyc01 mutant flies collected at the indicated times and probed with guinea pig VRI antisera. The lower blot is a shorter exposure of the upper blot. VRI, VRI protein; VRI-PO3, dephosphorylated VRI; asterisk, nonspecific band; ZT, Zeitgeber time. (B) Western blot analysis of total protein extracted from heads of wild-type flies collected at 2 hr intervals during an LD 12:12 cycle and probed with guinea pig anti-VRI antiserum. Nonspecific bands are denoted as closed diamond and asterisk. (C) The relative OD of VRI protein abundance (ratio of VRI/asterisk, closed triangle) was fit using a fifth order polynomial (thick black line) and plotted against a composite Clk mRNA (thick gray line) abundance curve. In cyc01 mutants, VRI protein levels are constitutively low (thin solid line) and Clk mRNA transcripts are constitutively high (thin dashed line). Neuron 2003 37, 249-261DOI: (10.1016/S0896-6273(03)00002-3)

Figure 2 The Clk Regulatory Region Contains Multiple E4BP4-Consensus Sites (A) Schematic of the Clk genomic locus. +1, the major transcription start site; ATG, start of translation; dashed arrow, minor transcription start site; white boxes, untranslated exon from major transcript; black boxes, translated exons; bent lines, introns; straight lines, untranscribed sequence. The arrow-tipped line represents Clk genomic sequences contained in the Clk8.0-Gal4 transgene. Asterisks show regions that have high similarity (9/10 or 10/10 nucleotide identity) with the consensus E4BP4 binding site in either orientation. Numbers above asterisks are based on the location of the +1 sense residue of the 3′ half site in relation to the major transcription start site and denote two sites used for EMSA analysis. (B) An 8.0 kb Clk genomic region contains regulatory elements sufficient for circadian transcription. Quantification of Gal4 mRNA levels from wild-type flies containing the Clk8.0-Gal4 transgene. RNase protection assays were performed on head mRNA from flies collected at the indicated times under LD conditions. The peak Gal4 mRNA level (black line) was normalized to 1.0. Gal4 mRNA levels are based on six independent experiments using three independent lines and plotted along with a composite Clk mRNA curve (gray line) from wild-type flies (Williams and Sehgal, 2001). White and black boxes represent times when lights were on or off, respectively. Neuron 2003 37, 249-261DOI: (10.1016/S0896-6273(03)00002-3)

Figure 3 VRI Binds E4BP4-Consensus Sequences from the Clk Locus EMSAs showing VRI-specific binding of E4BP4-consensus sites from the Clk locus. Each set of experiments was repeated at least three times with similar results. (A) The proteins used were wild-type VRI generated in rabbit reticulocyte lysates (+); basic domain-deleted version of VRI generated in rabbit reticulocyte lysate (Δb); and unprimed lysate (−). The probes were wild-type (+) or mutant (m1) −6651Clk and wild-type (+) or mutant (m1) −209Clk. Competition experiments were done with >100-fold excess of unlabeled wild-type (+) or mutant (m1) −6651Clk, wild-type (+) or mutant (m1) −209Clk, or no competitor (−). VRI shift, VRI shifted product; Free probe, unbound probe; asterisk, probe remaining in the well. (B) EMSAs showing VRI antisera block binding of VRI. Baculovirus-generated VRI was purified and used in EMSAs with the −6651Clk probe. Ab, VRI antiserum; PI, preimmune serum. Symbols are as described in (A). (C) EMSAs defining sites important for VRI binding. Three additional mutants (m2, m3, and m4) were generated in the context of the −6651Clk site to use as probes for VRI binding. Symbols are as described in (A). Neuron 2003 37, 249-261DOI: (10.1016/S0896-6273(03)00002-3)

Figure 4 Overexpression of VRI In Vivo Causes a Reduction in Clk mRNA Levels (A) Overexpression of VRI in wild-type flies. Wild-type flies, flies overexpressing VRI (VRI OX), and heterozygous vri mutant flies (vri2/+) were entrained for at least 3 days to an LD 12:12 cycle and collected at the indicated time points. Western blots showing relative VRI levels (left) and quantified ribonuclease protection assays (RPAs) showing the relative levels of Clk mRNA (Clk/rp49 ratio ± SEM) plotted against time of day (ZT) (right) are shown. Black line, Clk mRNA levels in flies overexpressing VRI; gray line, Clk mRNA in wild-type flies; dashed line, Clk mRNA in vri2/+ flies. RNA levels for each genotype are relative to the Clk mRNA peak at ZT 01 in wild-type flies, which was set to 1.0. Black asterisk, Clk mRNA values that are significantly different (p < 0.0004) from those at ZT 01 in vri2/+ flies; gray asterisk, Clk mRNA values that are significantly different (p < 0.0002) from those at ZT 01 in wild-type flies; open diamond, Clk mRNA values that are significantly different (p < 0.0002) from those at ZT 01 in VRI OX flies; dagger, Clk mRNA values in VRI OX and vri2/+ flies that are significantly different (p < 0.0002) than the values for wild-type flies at the same time point; double dagger, Clk mRNA values in vri2/+ flies that are significantly different (p < 0.02) than the values for VRI OX flies at the same time point. The data represent six independent experiments for VRI OX flies, five independent experiments for wild-type, and four independent experiments for heterozygous vri mutant flies. (B) Overexpression of VRI in cyc01 flies. cyc01 flies containing a heat-inducible vri transgene (hs-vri) and cyc01 controls were reared in constant light (LL) at 22°C, heat shocked at 37°C for 30 min, and allowed to recover at 22°C in LL (upper Western blots and RPA quantification) or reared in constant light at 18°C, heat shocked at 37°C for 15 min, and allowed to recover at 18°C in LL (lower Western blots and RPA quantification). hs-vri;cyc01 (V) or cyc01 (−) flies were collected before heat induction (NO) or 3 hr, 4 hr, 5 hr, 6 hr, 7 hr, 8 hr, or 24 hr after heat induction and used to determine the levels of VRI. Wild-type (wt) flies were collected at ZT 3 and ZT 15 to compare levels of VRI. Two independent VRI time courses are shown for each heat shock regime. Total head mRNA was subjected to RPA using Clk and rp49 probes to determine relative levels of Clk mRNA. Quantification of Clk mRNA levels is shown for heat-induced hs-vri;cyc01 flies normalized to the level of Clk mRNA in heat-shocked cyc01 flies at each time point. In the 30 min heat shock regime, the no, 3 hr, 4 hr, and 5 hr time points represent three independent experiments (error bars = SEM), and the 6 hr, 7 hr, and 8 hr time points represent two independent experiments (error bars = range). In the 15 min heat shock regime, the no. 3 hr, 4 hr, and 5 hr time points represent three independent experiments (error bars = SEM) and the 6 hr, 7 hr, and 8 hr time points represent two independent experiments (error bars = range). Asterisk, Clk mRNA values for the time points that could be statistically analyzed (i.e., no, 3 hr, 4 hr, and 5 hr for the 30 min heat shock and no, 3 hr, 4 hr, 5 hr, and 24 hr for the 15 min heat shock) that were significantly different (p < 0.02) than the non-heat-induced controls. Neuron 2003 37, 249-261DOI: (10.1016/S0896-6273(03)00002-3)

Figure 5 VRI Mediates cry mRNA Cycling (A) Schematic of the cry genomic region. Symbols are as described in Figure 2A, except the arrow-tipped line represents cryGal4 (Emery et al., 2000b). (B) A 6.0 kb cry genomic region contains regulatory elements sufficient for circadian transcription. Quantification of Gal4 (black line) or cry (gray line) mRNA levels from wild-type flies containing the cryGal4 transgene (Emery et al., 2000b). RNase protection assays were performed on head mRNA from flies collected at the indicated times under LD conditions. The peak Gal4 and cry mRNA levels were normalized to 1.0. White and black boxes represent times when lights were on or off, respectively. Two independent experiments are shown. (C) VRI binds E4BP4-consensus sites in the cry promoter. EMSA of competitions using 0.5 pmol of −6651Clk (+) or cry sites 1, 2, or 3. Symbols are as described in Figure 3. (D) Overexpression of VRI reduces cry mRNA in vivo. Flies were entrained and collected as in Figure 4A. RPAs were performed and plotted as in Figure 4A, except cry and rp49 probes were used. Black line, cry mRNA levels in flies overexpressing VRI; gray line, cry mRNA in wild-type flies. RNA levels are relative to cry mRNA at ZT 01 in wild-type flies, which was set to 1.0. Gray asterisk, cry mRNA values that are significantly different (p < 0.0006) from those at ZT 01 in wild-type flies; black asterisk, cry mRNA values that are significantly different (p < 0.03) from those at ZT 01 in VRI overexpression flies; dagger, cry mRNA values in VRI overexpression flies that are significantly different (p < 0.0005) than the values for wild-type flies at the same time point. The data represent six independent experiments for the VRI overexpression flies and five independent experiments for wild-type flies. Neuron 2003 37, 249-261DOI: (10.1016/S0896-6273(03)00002-3)

Figure 6 Model of the Interlocked Feedback Loop Mechanism in Drosophila (A) The per/tim loop (left) and dClk loop (right) are shown. In the per/tim loop, CLK-CYC heterodimers activate per and tim transcription via E boxes in the per and tim promoters. PER and/or TIM then bind CLK-CYC, repressing this activation and thus completing the per/tim negative feedback loop. In the Clk loop, CLK-CYC heterodimers activate vri transcription via E boxes in the vri promoter. VRI then binds VRI boxes in the Clk promoter and represses transcription, possibly in concert with a second repressor, thus completing the Clk negative feedback loop. PER/TIM positive regulation of Clk occurs indirectly by preventing CLK-CYC from activating vri. VRI also rhythmically represses VRI box-containing output genes whose mRNA transcripts cycle in phase with Clk (e.g., cry). Such genes may share a common activator with Clk, or they may be activated by independent factor(s). Filled arrows, transcriptional activation; open arrows, translation; bars, repression; gray arrow, output activation; gray bar, output repression. (B) Phases of clock gene product cycling. per, tim, and vri are activated by CLK-CYC in the early morning (ZT 04; blue line) and reach peak levels in the early evening (ZT 14). As per/tim/vri mRNA levels rise, Clk mRNA levels fall (black line) due to repression by rising levels of VRI (ZT 04, blue line). Unlike VRI, PER and TIM don't start accumulating until ZT 12 and reach high levels by ZT 18. Rising levels of PER-TIM bind CLK-CYC and prevent activation of per, tim, and vri, hence, per/tim/vri transcripts begin to decline during the early/mid-night (ZT 16; blue line). Declining levels of vri mRNA and protein allow Clk transcription to resume (ZT 18; black line). CLK-CYC dimers then begin to accumulate but are sequestered by PER/TIM during the late night/early morning (ZT 21–ZT 02, pink line), thus preventing premature per/tim/vri activation and Clk repression. When PER/TIM falls to trough levels (ZT 04; pink line), CLK-CYC is released and activates per, tim, and vri, thereby starting the next cycle. Open bar, lights on; closed bar, lights off. Neuron 2003 37, 249-261DOI: (10.1016/S0896-6273(03)00002-3)